Staged Training Report ✓ Complete

Run ID: decoder_only_embed384_depth10
Generated: 2026-02-25 22:57:24
Stages Completed: 1
Total Elapsed Time: 00:04:20

Configuration

No config defaults changed since last commit.

All Staged Training Parameters (61 parameters)

Parameter	Value
`total_samples`	10000000
`batch_size`	8
`stage_samples_multiplier`	100000000000
`update_interval`	250
`window_size`	100
`num_best_models_to_keep`	1
`sampling_mode`	Loss-weighted
`loss_weight_temperature`	0.5
`loss_weight_refresh_interval`	50
`stop_on_divergence`	True
`divergence_gap`	0.002
`divergence_ratio`	1.5
`divergence_patience`	50
`divergence_min_updates`	10
`val_spike_threshold`	2.0
`val_spike_window`	15
`val_spike_frequency`	0.75
`val_plateau_patience`	250
`val_plateau_min_delta`	0.0001
`custom_lr`	0.0001
`disable_lr_scaling`	True
`custom_warmup`	-1
`lr_min_ratio`	0.001
`resume_warmup_ratio`	0.05
`plateau_factor`	0.8
`plateau_patience`	15
`preserve_optimizer`	False
`preserve_scheduler`	True
`samples_mode`	Train additional samples
`num_random_obs_to_visualize`	2
`selected_frame_offset`	3
`runs_per_stage`	5
`serial_runs`	True
`clean_old_checkpoints`	True
`enable_baseline`	False
`baseline_runs_per_stage`	1
`run_id`	decoder_only_embed384_depth10
`seed`	None
`enable_wandb`	True
`wandb_project`	developmental-robot-movement
`lr_sweep.lr_min`	1e-07
`lr_sweep.lr_max`	0.01
`lr_sweep.phase_a_num_candidates`	5
`lr_sweep.phase_a_seeds`	1
`lr_sweep.phase_a_time_budget_min`	3.0
`lr_sweep.phase_a_survivor_count`	2
`lr_sweep.phase_b_seeds`	3
`lr_sweep.phase_b_time_budget_min`	10.0
`lr_sweep.ranking_metric`	median_best_val
`lr_sweep.min_samples_before_timeout`	1000
`lr_sweep.min_evals_before_stop`	5
`lr_sweep.save_sweep_state`	True
`plateau_sweep.enabled`	True
`plateau_sweep.plateau_ema_alpha`	0.85
`plateau_sweep.plateau_improvement_threshold`	0.0015
`plateau_sweep.plateau_patience`	25
`plateau_sweep.cooldown_updates`	5
`plateau_sweep.max_sweeps_per_stage`	2
`plateau_sweep.min_sweep_improvement`	0.0
`initial_sweep_enabled`	True
`stage_time_budget_min`	180.0

World Model Architecture (config.py)

Parameter	Value
`AUTOENCODER_LR`	0.0002
`BATCH_SIZE`	1
`CANVAS_HISTORY_SIZE`	3
`DECODER_DEPTH`	8
`DECODER_EMBED_DIM`	384
`DECODER_NUM_HEADS`	6
`DECODER_ONLY_DEPTH`	10
`EMBED_DIM`	384
`ENCODER_DEPTH`	5
`FOCAL_BETA`	5
`FOCAL_LOSS_ALPHA`	0.1
`FRAME_SIZE`	(224, 224)
`GRADIO_UPDATE_INTERVAL`	1
`LR_MIN_RATIO`	0.001
`MODEL_TYPE`	decoder_only
`NUM_HEADS`	6
`PATCH_SIZE`	16
`PERCEPTUAL_LOSS_WEIGHT`	0
`SEPARATOR_WIDTH`	16
`WARMUP_STEPS`	1000
`WEIGHT_DECAY`	0.01
`MASK_RATIO_MIN`	1
`MASK_RATIO_MAX`	1
`TRAIN_MASK_RATIO_MIN`	0.5
`TRAIN_MASK_RATIO_MAX`	1.0

Timing Summary

Stage	Plateau Sweeps	Sweep Time	Training Time	Stage Total
Stage 1	0	00:00:00	00:00:06	00:00:06
TOTAL	0	00:00:00	00:00:06	00:00:06

Initial LR Sweep: Stage 1: selected LR 1.00e-07 in 00:01:10

Stage Results

Stage	Best Loss	Stop Reason	Samples Trained	Time	Sweeps	LR (Initial→Final)
Stage 1	N/A	max_samples	0	00:00:06	0	1.0e-07

Total Plateau Sweeps: 0

Stop Reason Breakdown

max_samples: 1 stages

Multi-Run Statistics

Stage 1 (5 runs)

Run	Best Loss	Stop Reason	Time	Selected
1	N/A	max_samples	00:00:06	✓
2	N/A	max_samples	00:00:04
3	N/A	max_samples	00:00:04
4	N/A	max_samples	00:00:04
5	N/A	max_samples	00:00:04

Best Checkpoint

Name: stage1_run1_final.pth
Stage: 1
Hybrid Loss (full session): 5.501557

Learning Rate Timeline with Plateau Sweeps

Stage Progression

Stage	Orig Loss	Train Loss	Time	Samples	Stop Reason
1 ⭐	5.501557	N/A	00:00:06	0	max_samples

Hybrid Loss Over Original Session (per Stage)

Stage 1 (Best) - Hybrid Loss: 5.501557

Sample Counts

Cumulative Across All Stages

No cumulative sample count data available

Per Stage

No per-stage sample count graphs available

Best Checkpoint Inference

Selected Frame 3

Action 0

Action 1

Action 2

Random Observations

Observation 147

Action 0

Action 1

Action 2

Observation 769

Action 0

Action 1

Action 2

Individual Stage Reports

Stage 1 Run 1 (Staged) (orig loss: 5.501557) ⭐